c/mirror_zfs
1
0
Fork 0
mirror of https://git.proxmox.com/git/mirror_zfs.git synced 2024-11-17 10:01:01 +03:00
Code Issues Packages Projects Releases Wiki Activity

Do not hold spa_config in ZIL while blocked on IO

Browse Source 
Otherwise, we can get a deadlock that looks like this:

1. fsync() grabs spa_config_enter(zilog->zl_spa, SCL_STATE, lwb,
RW_READER) as part of zil_lwb_write_issue() . It then blocks on the
txg_sync when a flush fails from a drive power cycling.

2. The txg_sync then blocks on the pool suspending due to the loss of
too many disks.

3. zpool clear then blocks on spa_config_enter(spa, SCL_STATE |
SCL_L2ARC | SCL_ZIO, spa, RW_WRITER)  because it is a writer.

The disks cannot be brought online due to fsync() holding that lock and
the user gets upset since fsync() is uninterruptibly blocked inside the
kernel.

We need to grab the lock for vdev_lookup_top(), but we do not need to
hold it while there is outstanding IO.

This fixes a regression introduced by
1ce23dcaff.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Richard Yao <richard.yao@klarasystems.com>
Sponsored-By: Wasabi Technology, Inc.
Closes #14519
This commit is contained in:
Richard Yao 2023-03-07 19:12:28 -05:00 committed by GitHub
parent 1f196e3107
commit 7d638df09b
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 2 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
module/zfs/zil.c
View File

@ -1287,8 +1287,6 @@ zil_lwb_flush_vdevs_done(zio_t *zio)
itx_t *itx; itx_t *itx;
uint64_t txg; uint64_t txg;
spa_config_exit(zilog->zl_spa, SCL_STATE, lwb);
zio_buf_free(lwb->lwb_buf, lwb->lwb_sz); zio_buf_free(lwb->lwb_buf, lwb->lwb_sz);
mutex_enter(&zilog->zl_lock); mutex_enter(&zilog->zl_lock);
@ -1427,8 +1425,6 @@ zil_lwb_write_done(zio_t *zio)
zil_vdev_node_t *zv; zil_vdev_node_t *zv;
lwb_t *nlwb; lwb_t *nlwb;
ASSERT3S(spa_config_held(spa, SCL_STATE, RW_READER), !=, 0);
ASSERT(BP_GET_COMPRESS(zio->io_bp) == ZIO_COMPRESS_OFF); ASSERT(BP_GET_COMPRESS(zio->io_bp) == ZIO_COMPRESS_OFF);
ASSERT(BP_GET_TYPE(zio->io_bp) == DMU_OT_INTENT_LOG); ASSERT(BP_GET_TYPE(zio->io_bp) == DMU_OT_INTENT_LOG);
ASSERT(BP_GET_LEVEL(zio->io_bp) == 0); ASSERT(BP_GET_LEVEL(zio->io_bp) == 0);
@ -1490,6 +1486,7 @@ zil_lwb_write_done(zio_t *zio)
return; return;
} }
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
while ((zv = avl_destroy_nodes(t, &cookie)) != NULL) { while ((zv = avl_destroy_nodes(t, &cookie)) != NULL) {
vdev_t *vd = vdev_lookup_top(spa, zv->zv_vdev); vdev_t *vd = vdev_lookup_top(spa, zv->zv_vdev);
if (vd != NULL) { if (vd != NULL) {
@ -1505,6 +1502,7 @@ zil_lwb_write_done(zio_t *zio)
} }
kmem_free(zv, sizeof (*zv)); kmem_free(zv, sizeof (*zv));
} }
spa_config_exit(spa, SCL_STATE, FTAG);
} }
static void static void
@ -1783,8 +1781,6 @@ zil_lwb_write_issue(zilog_t *zilog, lwb_t *lwb)
*/ */
memset(lwb->lwb_buf + lwb->lwb_nused, 0, wsz - lwb->lwb_nused); memset(lwb->lwb_buf + lwb->lwb_nused, 0, wsz - lwb->lwb_nused);
spa_config_enter(zilog->zl_spa, SCL_STATE, lwb, RW_READER);
zil_lwb_add_block(lwb, &lwb->lwb_blk); zil_lwb_add_block(lwb, &lwb->lwb_blk);
lwb->lwb_issued_timestamp = gethrtime(); lwb->lwb_issued_timestamp = gethrtime();
lwb->lwb_state = LWB_STATE_ISSUED; lwb->lwb_state = LWB_STATE_ISSUED;